Developing sensitive and valid indicators of Alzheimer's Disease (AD), dementia and cognitive function in low literacy and low income countries is challenging and requires creative approaches to complement standard neuropsychological batteries. In this proposal, we aim to incorporate an innovative approach to assessing cognition and to validate it against standard cognitive batteries, medical and informant assessments, brain function assessed from MRI's and genetic analyses of APOE in 3 HRS sister countries: South Africa (HAALSI), United States (HRS/COG), and Ireland (TILDA). South Africa has a low literacy rate in older people, varying degrees of use of written language and different cultural cues compared to Western industrialized countries. In order to assess cognition in this setting, we assess cognitive function leveraging non-verbal and minimally numeric assessments for a range of cognitive domains. Through administration via tablets, responses are recorded and scored immediately. The tablet- based assessments will be validated in relation to the Harmonized Cognitive Assessment Protocol (HCAP) developed by HRS across HAALSI, TILDA and HRS/COG (Aim 1). Additional information from MRI's and genetic analyses increase validity. Clinical and informant interviews are added to develop an AD and dementia risk score (Aim 2). We conduct the validations and harmonization in age and education stratified samples of 300 men and women 50 -79 years old in each of the three cohorts (n=900). Novel `next generation' instruments are being developed and we now need information regarding their validity compared to standard cognitive assessments. Additional psychometric evaluation in collaboration with PITCH (R01AG051170, Jones PI) will aid in refining these measures. Developing cross-walks between these measures and other commonly used indices is of special importance in the set of sister studies related to the Health and Retirement Study where the explicit goals are to permit comparisons across countries, especially with regard to AD and dementia. TILDA and HAALSI offer specific advantages in calibration because they have had rich additional neuropsychological and health data from earlier waves across an educationally diverse older population. We have two aims in our proposal: Aim 1. To assess the validity of the Oxford based language controlled cognitive measures in relation to the HCAP measures using identical platforms for HRS/COG, TILDA and HAALSI 1.a. to evaluate content, criterion and construct validity of Oxford assessments against 1.b. to develop a refined battery of items based on the psychometric analyses 1.c. to develop a shorter and harmonized set of items that have good psychometric properties Aim 2. To develop a common dementia risk score in each country based on the additional information gathered from proxy/informants and a clinical evaluation. This risk score will be harmonized and validated.